When it's time to move your business off Excel: 7 signs spreadsheets are holding you back

It's time to leave Excel when the spreadsheet stops being a draft and becomes the company's database. The statistics show how risky that is: in Professor Ray Panko's review of 13 studies, 94% of the audited working spreadsheets contained at least one error. While the business is small, you can live with that. When your accounting rests on spreadsheets, every error costs money.

Excel is not the enemy. It is the best tool for a fast start. The problem is that the business grows and the spreadsheets do not: they do not let ten people work at once, they do not validate data on entry, and they stay silent when the numbers disagree. Below are seven signs you have already outgrown spreadsheets, and how to move on without stopping work.

7 signs spreadsheets are already costing you money

There is one main sign: the same number is known differently across the company. Everything else is detail. If you recognize yourself in at least three of the seven points, spreadsheets are already working against you — the bill just arrives late and is spread across the salaries of people who reconcile data by hand, and across decisions made on stale numbers.

1. The same number is typed in by hand in two or more places — and it differs in each.
2. There is an employee whose job is half "reconciling" spreadsheets.
3. The report for the owner takes more than a day to prepare and is out of date by the time it is ready.
4. One person leaving paralyzes the process: only they know how the file works.
5. Errors surface only a month later — during reconciliation with the bank or the warehouse, when fixing them is already expensive.
6. Managers keep their own spreadsheets, and nobody has the full picture of customers.
7. Any new question ("what's the margin on this line of business?") turns into a week-long project.

Recognize yourself in three points or more and you are paying the "Excel tax" every month. It is not a separate line in the budget, so it is easy to miss: it is paid in employees' time, in accounting errors, in decisions made on last quarter's data.

What is the "Excel tax" and how do you calculate it?

The "Excel tax" is the hidden cost of manual accounting — the amount a company pays every month for keeping its data in spreadsheets. It is calculated in a straightforward way: take the hours employees spend consolidating and reconciling, multiply by their rate and add the cost of errors — the discrepancies that surface a month later. It usually comes out to one full salary.

The formula is the same as for any routine work: frequency of the operation × time × rate + cost of errors. A bank reconciliation that takes an accountant two days a month is a month of work over a year. How to calculate the cost of routine work area by area, and where to start, is covered in the hub article “Business automation: where to start”.

How is a single integrated system different from spreadsheets?

A single integrated system means an operation is entered once, at the moment it happens, and from there flows into accounting, reporting and analytics on its own. A spreadsheet stores numbers that someone typed into it; an integrated system records an event and validates it on entry. The difference shows up in four things: where the number is born, who validates it, when a discrepancy surfaces, and who is responsible for making the numbers add up.

The last row is about the owner no longer depending on someone else's memory. Which numbers to keep in front of you every day to see the state of the business as it is right now — that's covered in the piece on the owner's dashboard.

Is it true that companies lose real money because of Excel?

Yes, and not only small companies. Some of the world's largest organizations have lost billions and months of work on spreadsheets. The cause is always the same: a spreadsheet does exactly what it is asked to do, including quietly dropping data beyond the visible range. Four public cases show how it ends when a file becomes the accounting system.

• The UK undercounted 16,000 COVID cases. Public Health England transferred test results in the legacy .xls format with its row limit — some records simply never made it into the count, and contact tracing notifications were never sent (The Register, October 2020).
• JPMorgan — 6.2 billion dollars. The risk valuation model ran on a chain of spreadsheets filled in by manual copy-paste; an error during a transfer understated the risk and became one of the causes of the "London Whale" episode (Full Stack Modeller).
• Williams F1 was building a car from a 20,000-row spreadsheet. The team's new principal discovered that the entire parts list for the car was kept in a single file without prices, stock levels or delivery dates — "impossible to navigate or update" (Autosport, 2023).
• Scientists had to rename 27 genes. Excel kept turning symbols like MARCH1 into dates and corrupting scientific data, so the nomenclature committee changed the gene names themselves to fit the spreadsheet's behavior (The Verge, 2020).

What all four have in common: the spreadsheet outgrew the role of a draft but stayed in the place of a database. In our custom development work we start a project at exactly this point: we find where accounting rests on a file and one person, and move it into a system that validates the data itself.

Why doesn't an off-the-shelf product often take root?

An off-the-shelf product fails to take root when its logic doesn't match the company's process. The typical scenario: a ready-made system is bought, processes are bent to fit it for three months, and then the managers quietly go back to Excel — "it's faster this way". The system stays paid for and empty. The one that takes root is the one that mirrors the real process and makes life easier for the person entering the data.

That is why the first step of the move is mapping the processes: where data is born, who enters it, who needs it downstream. The choice of platform comes second. Often a separate system isn't needed at all: a standard area is covered by a ready solution from our catalog, and the non-standard core is built out with custom development.

Can you just replace the spreadsheet with a quick utility?

Sometimes yes — if the spreadsheet is a one-off and losing it wouldn't hurt. A converter, a price-list parser, a one-time calculation can be put together in an evening even without a programmer, by describing the task to an AI in plain words. That's vibe coding, and for draft-grade utilities it works very well. But as soon as money, customers or accounting depend on that spreadsheet, the quick replacement becomes a new point of failure.

The line runs along the cost of an error. While a failure costs five minutes of inconvenience, a quick utility is justified and saves weeks. When a failure means a lost order or a discrepancy with the bank, you need a system with validation on entry, contractual responsibility and support after delivery. Vibe-coded utilities are a good fit for exactly the transition zone: while you're deciding what to do about the big spreadsheet, they take away part of the manual pain.

How do you move off Excel without stopping the business?

The move happens in short stages, each of which delivers value on its own — no "big bang" where everything changes in a single day. First the most painful area is automated, the old process runs in parallel for a while for cross-checking, and only then are the adjacent areas connected. The first working area appears within weeks, long before the end of the project.

1. Automate the most expensive area — inventory or payment reconciliation: the one where the "Excel tax" is highest.
2. Keep the old process running in parallel while the numbers are cross-checked and trust in the system grows.
3. Migrate historical data: clean up duplicates, bring reference lists to a single format — sales and customer history is not lost.
4. Connect the adjacent areas: finance, payroll, reporting.
5. Leave Excel where it is good — one-off calculations and drafts, not the company's database.

In our case with a retail chain, this kind of move removed manual inventory accounting entirely: a Telegram bot keeps the records, and the errors that used to be found once a month simply don't occur any more. A separate task is connecting what already works: 1C — the accounting and ERP platform most Russian companies run on — in the accounting department, the CRM in sales, the bank. How systems are brought together into one integrated setup without replacing everything with a "mega-system" is covered in the article on integrations.

Where do you start the move?

Start with the one spreadsheet that hurts the most. Calculate the "Excel tax" on it: hours spent consolidating × rate + cost of errors per month. That number is the budget you are already spending, just invisibly. Then compare it with the cost of automating that area and you get a payback period — which turns the decision to leave spreadsheets into an ordinary investment decision.

If you'd rather not do the math yourself — tell us about your task. We'll map the processes, find where accounting rests on a file and one person, calculate the cost of the routine work and propose a plan with a target expressed in money. The data stays in Russia, and support continues after delivery.

Sources

• Ray Panko: a review of research on spreadsheet errors (Spreadsheet Errors: What We Know)
• The Register: England undercounted 16,000 COVID cases because of Excel's row limit, October 2020
• Full Stack Modeller: an analysis of the Excel model error in JPMorgan's "London Whale" story
• Autosport: Williams dropped its 20,000-row Excel spreadsheet for building the car, 2023
• The Verge: scientists renamed 27 genes so that Excel wouldn't read them as dates, 2020